At the Southwest Healthcare Innovation Summit, being held at the Marriott Medical Center in downtown Houston and sponsored by Healthcare Innovation, Stan Huff, M.D., Chief Medical Informatics Officer of the 23-hospital, 215-clinic, Salt Lake City-based Intermountain Healthcare, on November 14, presented a strong vision of interoperability in his keynote presentation, “Playing with FHIR! How Intermountain Healthcare Is Driving Change in Clinical Interoperability.”
Why have he and his colleagues at Intermountain focused on interoperability? As he noted in a slide in his presentation, simply this: “To help people live the healthiest lives possible.”
Intermountain, as Dr. Huff noted, has a history of work towards interoperability going all the way back to Dr. Homer Warner, and the early 1960s, even before computers were part of patient care delivery.
Further, as a slide in his presentation noted, “Intermountain can only provide the highest quality, lowest cost health care with the use of advanced clinical decision support systems integrated into frontline clinical workflow.”
Further, Huff noted, clinical practice has actually become more complex, thus potentially opening itself up to more medical errors and issues, than ever before. As it put it, “In the past, medicine was simple, relatively safe, and ineffective. Now, it’s costly, complex, and potentially dangerous.”
And he shared a tragic case study, involving a young woman named Eileen, an informaticist from Singapore, who had collapsed in her hotel in San Antonio, in 2012. After collapsing, she was rushed to the emergency department of a local hospital, and was intensively treated for what turned out to be septic shock, though the sepsis was not identified early enough in her emergency treatment. Tragically, the physicians and nurses were unable to save Eileen.
As Huff noted, “Clinical decision support should have saved her.” Speaking of Intermountain’s intensive work on sepsis treatment, he noted that “We went from a mortality rate of 14 percent to 7 percent. There are clinical decision support tools in most EHRs [electronic health records], and work is being done in most organizations. But most people are being cared for in 200- and 300-bed community hospitals,” not in the academic medical center-based health systems in which so much progress is being made in that area. “That’s why we’re trying to be a model at Intermountain,” he said.
“What’s startling,” Huff added, “is that Eileen’s case is not unusual. The number of people in the United states who die every year from preventable medical errors is 251,454. Meanwhile, 40,000 die in auto accidents every year. This is six times as many people.” Further, he added, “Looking at the opioid crisis, 100-150 people die every day from opioid overdoses, while fully 700 die from preventable medical errors.” With such high rates of deaths due to medical errors, why is there no nationwide outcry? “It’s diffuse, and some people are in denial; they don’t believe the numbers. I believe the numbers, not only because of people like Eileen, but because of my practice as a physician,” he said, vowing that he would continue to advocate for intensive work to bring down medical errors healthcare system-wide. “Not only are we hurting and killing patients unintentionally; but there’s a real impact on healthcare workers as well; it can have tragic effects on them,” he said.
And he quoted two historic healthcare system leaders, David Eddy, M.D. and Clement J. McDonald, M.D. Dr. Eddy has said, “The complexity of modern medicine exceeds the inherent limitations of the unaided human mind.” And Dr. McDonald has said, “[M]an is not predictable. There are limits to man’s capabilities as an information processor that assure the occurrence of random errors in his activities.”
He added that “People, even when they know the right thing to do, don’t always do it; we’re not always perfect information processors. Doctors have a lot going on.”
Huff and his colleagues have created at least 1,600 decision support rules or modules at Intermountain Healthcare. Among those many are decision support modules for ventilator weaning, nosocomial infection monitoring, MRSA monitoring and control, prevention of deep venous thrombosis, diabetic care, pre-operative antibiotics, ICU glucose protocols, ventilator disconnection, and many, many others.
Speaking of those 1,600 rules and modules, he said, “We have picked the low-hanging fruit.” In fact, he said, “There is a need to have 10,000-plus decision support rules or modules. What’s more, he said, “There is no path to get from 1,600 to 10,000 unless we fundamentally change the ecosystem.” The fundamental challenge? “There is no scalable path [for the sharing of large numbers of CDS tools] from the leading institutions to community hospitals.”
Importantly, Huff said, “We can focus on the most obvious things, the most devastating things, the most costly things. You go down the list of conditions like asthma, myasthenia gravis, etc., and we’re not doing anything for those patients. We’ve been focused on the most common conditions. But we have to help everyone. And we need to be able to share knowledge, not in journals, but through apps that you can download from the app store to your device.”
And that situation speaks to a core challenge. “We can’t do that now,” Huff said, referencing the broad downloading of clinical apps, “because everybody’s system is different—not only in terms of different EHRs, but even in terms of different instances of the same EHRs from the same vendors. And that means that every useful clinical decision support system we create, has to be recreated in every new system. And that’s not scalable, and it won’t get to the 300-bed hospital in San Antonio where Eileen was seen. And that’s where interoperability comes in. And interoperability isn’t a single thing. It’s many things.”
Huff said that a series of steps from version to of the HL7 standard to the FHIR (fast healthcare interoperability resources), to “FHIR according to the Argonaut [Project] rules,” is moving the industry forward, eventually leading to true interoperability. “
“We need to move towards standard representations of data, across all information systems,” Huff said. “Then you can write applications, and those applications don’t know whether they’re talking to Cerner or Epic or Allscripts or some other system; they’re communicating through a FHIR API or other systems. So that you can literally load it in the same way you can load it onto an iPhone. That’s what we’re striving for.” Inevitably, in that, he said, “We’re in the early stages” as an industry.
It will be important, he said, to focus on the conditions and situations involving the greatest initial potential. “The management of acute community-acquire pneumonia, of glucose in the ICU; those are the kinds of things that will have an impact on mortality and on decreasing the cost of healthcare.”
With regard to investing in the future, Intermountain is one of the patient care organizations that have helped to create a not-for-profit company called Logica. Intermountain, the Veterans Administration healthcare system, LSU (Louisiana State University) Health, and others, have come together to create Logica as a not-for-profit 501c3, designed to create a vibrant, open platform for the development of interoperable applications, content and services. With a $300,000 annual budget, he said, “It really is a volunteer community.”
What does Logica do? Among other things: it is an open shared repository of detailed clinical models, licensed, and free for use by everyone. It involves the conversion of Intermountain models to CIMI (Clinical Information Modeling Initiative) models and FHIR profiles, and platform/SOA architecture, among numerous other elements.
With regard to standards-based healthcare interoperability,” Huff said, “I talk about these issues at national and international meetings. And I come home and we try to implement the things I talk about, and I find that Intermountain, with our 39,000 employees—there are roughly 500 IS employees who are creating applications. When we look at our own shop, they’re still doing this work in silos, focusing on particular programs or services who needed specific applications. We found that Intermountain Healthcare was not even doing that consistently within our own organization”—meaning that process change needs to take place and is taking place, at Intermountain itself.
“What’s being proposed,” Huff said, “is that with regard to lab instruments and test kits, that the manufacturers, who are the experts on what their instrument or test kit measures—what if those vendors would identify the LOINC code for that particular item? then it would not require mapping, just a lookup.”
Several critical success factors will be involved in the ultimate shift to true interoperability. The first will be clinical engagement and clinician ownership of process. The second will be technical tools, among them data; services (HL7, FHIR, SOA); knowledge representation; API adoption; and implementation. In addition, the healthcare industry needs a policy and legal framework to enable information-sharing. And it will require a business case, and financial incentives.
Huff said he sees a number of developments as exciting, including the Apple Health initiative; Google’s cloud healthcare API; Amazon’s FHIR serverless architecture, and Microsoft’s Azure API for FHIR.
“You can exchange data with CCDA, but it’s harder because of the flexibility with which people can represent their data,” he said, adding that “You need coded and structured data whose origins you know. And that data is now residing on my iPhone.”
And he concluded with two quotes from the great Catalan-Spanish architect Antoni Gaudí, whose masterpiece, the church of the Sagrada Familia in Barcelona, was his masterpiece. “What must be always preserved is the spirit of the work; its life will depend on the generations that transmit this spirit and bring it to life,” Gaudí said. The architect also said, “To do things right, first you need love, then technique.” “Interoperability won’t be solved in six months. We can make significant progress in five years and even more significant progress in ten years,” Huff said. “We have to be involved in this together. Over time, we’ll decreases the 250,000 people a year who are dying of avoidable medical errors. It’s absolutely worth it to do the hard parts.”
